I think one problem is when people talk about RPG theory there's this implication that a potential theory exists that would cover all play styles and game types, when there's no guarantee that that's the case. Tactical dungeon-raiding games and, say, relationship-focused storytelling games are so completely different in their inputs, outputs, and goals that it seems weird to even call them the same sort of activity.
Each style of play needs a different theoretical framework. I've only seen theory discussions be productive when they are very tightly focused on "how do I make this one particular style of play better?"